Do fanless computers typically have higher CPU temperatures?

ixfd64 · 2018-05-14, 22:56

I recently got access to an an embedded system with a nice quad-core Skylake processor. The computer has no fans and relies on external heat sinks for cooling. However, Open Hardware Monitor shows that the idle temperatures are around 48° C - or about 10 degrees above average. This machine has very little previous use.

So I'm curious: is it normal for fanless computers to have higher CPU temperatures? And is it safe to run Prime95 on embedded computers in general?

Thanks!

masser · 2018-05-14, 23:51

Quote:

Originally Posted by ixfd64

I recently got access to an an embedded system with a nice quad-core Skylake processor. The computer has no fans and relies on external heat sinks for cooling. However, OpenHardwareMonitor shows that the idle temperatures are around 48 ° C - or about 10 degrees above average. This machine has very little previous use.

So I'm curious: is it normal for fanless computers to have higher CPU temperatures?

Yes.

Quote:

Originally Posted by ixfd64

And is it safe to run Prime95 on embedded computers in general?

Thanks!

On a relatively newer processor, like skylake, it is probably safe; however, the chip will throttle as it approaches higher temperatures with Prime95 use. You will probably be disappointed by the performance.

On older processors in fanless systems, it might not be safe.

Xyzzy · 2018-05-15, 02:30

We have a fanless computer doing "PRP-CF" work.

It is an Intel Celeron N3060.

It does about 2GHz-Days/Day.

It runs lukewarm but since it is so slow we suppose that is to be expected.

Nick · 2018-05-15, 06:57

Quote:

Originally Posted by masser

...however, the chip will throttle as it approaches higher temperatures...

If the universe works the same way, then global warming will eventually slow down time.

mackerel · 2018-05-15, 08:02

How hot something runs depends on how much cooling potential there is relative to the power output. It's probably easiest to run P95 in stress test mode and see how hot it gets.

heliosh · 2018-05-15, 08:02

The problem most likely won't be the CPU, but the CPU heating the surrounding components. Especially electrolytic capacitors do age faster with higher temperatures. Make sure to allow a bit of convectional air flow through the case.

LaurV · 2018-05-15, 14:41

@OP, you may want to read this thread (around posts 8-10 or so)

ixfd64 · 2018-05-15, 15:58

Thanks, everyone. That's a lot of useful information!

kladner · 2018-05-15, 17:19

Quote:

The computer has no fans and relies on external heat sinks for cooling.

Can you gain anything by pointing a desk fan at the device?

chalsall · 2018-05-15, 18:14

Quote:

Originally Posted by ixfd64

Thanks, everyone. That's a lot of useful information!

A lot has already been given, but if I may share a bit more...

Compute consumes energy to decrease the entropy (read: get results). An enviable result of this is to increase the enthalpy (read: produce heat).

An important part of cooling is that the thermal delta is what matters most. This is why many Bitcoin miners are in Iceland, and why cars have radiators.

heliosh · 2018-05-15, 19:19

Also a hot CPU appears to run slower than a cold one. Not only because of said reason (thermal throttling), but because of general relativity.

A hot CPU contains more thermal energy, thus is heavier than a cold one. Because it has more mass, the time appears to run slower.
(But I guess it makes a smaller difference than one clock cycle in a billion years)

2018-05-14, 22:56	#1
ixfd64 Bemusing Prompter "Danny" Dec 2002 California 7⁴ Posts	Do fanless computers typically have higher CPU temperatures? I recently got access to an an embedded system with a nice quad-core Skylake processor. The computer has no fans and relies on external heat sinks for cooling. However, Open Hardware Monitor shows that the idle temperatures are around 48° C - or about 10 degrees above average. This machine has very little previous use. So I'm curious: is it normal for fanless computers to have higher CPU temperatures? And is it safe to run Prime95 on embedded computers in general? Thanks! Last fiddled with by ixfd64 on 2018-05-15 at 22:06

2018-05-15, 14:41	#7
LaurV Romulan Interpreter Jun 2011 Thailand 9664₁₀ Posts	@OP, you may want to read this thread (around posts 8-10 or so) Last fiddled with by LaurV on 2018-05-15 at 14:42

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
i7-8550U temperatures	lycorn	Hardware	6	2018-01-14 22:07
GTX 580 temperatures	Mark Rose	GPU Computing	7	2014-09-03 16:48
Worried about my temperatures..	Unregistered	Information & Answers	25	2011-12-30 21:42
What CPU Temperatures Are Acceptable?	Rodrigo	Information & Answers	9	2010-08-05 18:47
CPU Temperatures & Air Flow	ndpowell	Hardware	15	2005-06-25 15:19

2018-05-15, 02:30	#3
Xyzzy "Mike" Aug 2002 8245₁₀ Posts	We have a fanless computer doing "PRP-CF" work. It is an Intel Celeron N3060. It does about 2GHz-Days/Day. It runs lukewarm but since it is so slow we suppose that is to be expected.

2018-05-15, 08:02	#5
mackerel Feb 2016 UK 2²×109 Posts	How hot something runs depends on how much cooling potential there is relative to the power output. It's probably easiest to run P95 in stress test mode and see how hot it gets.

2018-05-15, 08:02	#6
heliosh Oct 2017 ++41 5³ Posts	The problem most likely won't be the CPU, but the CPU heating the surrounding components. Especially electrolytic capacitors do age faster with higher temperatures. Make sure to allow a bit of convectional air flow through the case.

2018-05-15, 15:58	#8
ixfd64 Bemusing Prompter "Danny" Dec 2002 California 2401₁₀ Posts	Thanks, everyone. That's a lot of useful information!

2018-05-15, 19:19	#11
heliosh Oct 2017 ++41 125₁₀ Posts	Also a hot CPU appears to run slower than a cold one. Not only because of said reason (thermal throttling), but because of general relativity. A hot CPU contains more thermal energy, thus is heavier than a cold one. Because it has more mass, the time appears to run slower. (But I guess it makes a smaller difference than one clock cycle in a billion years)